Phenology models are useful tools to study phenology shifts and their responses to climate change. Multiple factors including temperature, precipitation, photoperiod, insolation, and snow can affect the phenology of alpine grasslands on the Qinghai–Tibetan Plateau (QTP), but most models applied on the QTP have only considered the influences of temperature or temperature and precipitation. This study presents a multi-factor-driven phenology model, the Alpine Meadow Prognostic Phenology (AMPP) model, based on the Growing Season Index (GSI), to predict both leaf onset and offset dates of QTP alpine meadows at the community scale. Five factors including daily minimum air temperature, precipitation averaged over the previous month, photoperiod, global solar radiation, and snowfall were combined into an integrated index, the Alpine Meadow Growing Season Index, to quantify climatic limitations on foliar development of QTP alpine meadows. A case study was conducted using the observed leaf onset and offset dates of dominant species in QTP Kobresia meadows from 1989 to 2016. The root-mean-square errors (RMSEs) of modeled leaf onset and offset dates from the AMPP model were 6.9 d and 11.0 d, respectively, decreasing by 13.8%–48.9% and 7.6%–47.1% compared with the null model and seven other phenology models. The correlation coefficients between the predicted and observed leaf onset and offset dates were 0.75 and 0.34, respectively, higher than the 0.50–0.68 and −0.22–0.11 from other models. The RMSE ranges of predicted leaf onset and offset dates among three different sites were 2.9 d and 1.1 d, respectively, lower than or equal to the 3.6–12.4 d and the 1.1–6.9 d from other models. Results indicated that the AMPP model clearly improved the prediction accuracy and simulation of the interannual variability of leaf onset and offset dates and showed more robust simulations at different sites. Moreover, this model can be easily embedded into most ecosystem process models and applied to other alpine or subalpine grasslands once it has been adapted to their requirements.